文摘
Mining frequent itemsets can often generate a large number of frequent itemsets. Recent studies proposed mining itemset with the different types of constraint. The paper is to mine frequent itemsets, where a one: does not contain any item of C 0 or contains at least one item of C 0. The set of all those ones is partitioned into equivalence classes. Without loss of generality, we only investigate each class independently. One class is represented by a frequent closed set L and splits into two disjoint sub-classes. The first contains frequent itemsets that do not contain any item of C0. It is generated from the corresponding generators. The second includes in two subsets of the frequent itemsets coming from the generators containing in C 0, and the ones obtained by connecting each non-empty subset of L(C 0 with each element of the first.